 Despite the great success medicine has had in saving human lives, we have made very little progress in tackling chronic diseases, particularly brain conditions such as Alzheimer's, depression, autism and epilepsy that people may have to live with their whole lives. This is going to be a major grand challenge for medicine in the coming decade. Every four seconds there is a new case of dementia in the world. Every 40 seconds someone commits suicide and every 20 minutes more than one child is diagnosed with autism. These diseases are taking a silent toll on our lives and economies. The global financial cost of these diseases is estimated at $2.5 trillion per year and only expected to grow many fold as they are more widely acknowledged. The personal cost is incalculable as people dealing with them face mental, financial and social stigma throughout their lifespan. In all these diseases, the brain plays a central role in the prevention, progression and intervention. Why then isn't a neuroscientist standing in front of you? Why aren't there cures for Alzheimer's, depression, autism and epilepsy? Well, the brain of a tiny 1mm long worm known as C. elegans has 302 neurons with 7000 connections between them. Imagine how complicated then is the human brain where these numbers are in the hundreds of billions and trillions. This is where I come in, a researcher trained in machine learning. My team has developed tools to understand the complicated interplay between the intricacies of the brain and physical health by developing tools such as high dimensional functional regression that can analyze the activity and connections in nearly a million brain regions to infer the chronic basis of neural diseases. Machine learning tools such as these make the difference between a scientist 50 years ago looking at a single brain scan trying to figure out what has gone wrong with the brain and a computer able to process thousands of high resolution brain scans automatically determining the commonalities amongst them and the link to diseases. A specific focus of my research is on understanding which regions of the brain influence particular health outcomes. As an example, while studies have shown that stress negatively affects the brain, I have shown the reverse that there are certain regions in the brain which can tell us how reactive a person is to stress. Analytics such as these not only help us predict who is going to get a particular stress-induced disease and how it might progress but also pave the way for design of better brain-based intervention that can reduce the risk of those diseases. In another example, I have used the same machine learning techniques to vastly increase the number of brain connections that can be mapped to a person's characteristics such as age. By better understanding the normal aging of the brain, we can enable early diagnosis of age-related pathologies such as Alzheimer's disease. Until now, neuroscientists did not have such sophisticated tools that can leverage the full information in these data sets or learn from multiple types of modern brain data sets. And these have been the key limiting factors that are preventing the use of brain-based diagnosis in the practice of medicine. Machine learning tools that we are developing can not only learn from multiple types of brain scans, they can also incorporate additional factors such as genetics, lifestyle, personal history and social factors. And inarquably, combination of all of these data sets is necessary to enable accurate as well as personalized protection of brain diseases. Another dimension of my research focuses on designing intelligent algorithms that are feedback-driven and can automatically turn sensors on and off. These are electrodes in the brain. This helps us to design a very high-resolution yet low-power EEG system that can be used to monitor an epilepsy patient 24-7 even at home. And similar intelligent algorithms can also interact with a neurosurgeon, leveraging his insights while providing him guidance on inserting invasive probes to better localize the source of an epilepsy and plan for brain surgery. To sum up, machine learning is providing us powerful tools. And time is now ripe to unleash this power for understanding the complications of the brain and its intricate interplay with health. My question for you is, imagine you have access to a brain-based device that can measure stress levels or depression levels. What possibilities might this enable? Can we imagine putting pilots and combat troops through a stress exam or pregnant women to determine their risk of postpartum depression? What else? I'd love to hear your thoughts on this. Thank you.